As the US Food and Drug Administration (FDA) becomes more receptive to the use of real-world data (RWD) in drug and device application reviews, it is vital that sponsors ensure that data used to support clinical trials are high-quality and well-sourced.
The new guidance, which will allow for big data sources to be used without private, individual information, is currently in place for certain device submissions, but the FDA says it intends to update the guidance for drugs and biologics.
Go deeper with GlobalData
Discover B2B Marketing That Performs
Combine business intelligence and editorial excellence to reach engaged professionals across 36 leading media platforms.
Since 2016, 35 drugs, biologics, or vaccines have included real-world evidence (RWE) in their applications; however, its integration for devices has been more extensive, with over 250 premarket authorisations including RWE in the same period.
This policy change opens the door to using deidentified databases, including national cancer registries like the National Cancer Institute’s Surveillance, Epidemiology, and End Results, hospital systems databases, insurance claims databases, and electronic health records (EHRs). FDA Commissioner Marty Makary says this removes “unnecessary barriers” and will help provide treatments to patients faster.
Jen Lamppa, VP of Commercial Strategy at Inovalon, agrees with this sentiment but believes that sponsors need to understand what the FDA is allowing if they wish to successfully integrate this into their submissions.
This interview has been edited for length and clarity
US Tariffs are shifting - will you react or anticipate?
Don’t let policy changes catch you off guard. Stay proactive with real-time data and expert analysis.
By GlobalDataAbigail Beaney (AB): What does this change from the FDA actually mean in practice?
Jen Lamppa (JL): Previously, EHR-derived data has played a largely supportive role in regulatory submissions. The FDA primarily relied on RWD for post-marketing safety surveillance. It has only been selectively used in the past for effectiveness evaluations. Post-marketing RWD has mostly focused on safety, with EHR data being used to evaluate adverse events (AEs) or utilisation patterns in broader populations.. That’s where the landscape has shifted here.
With the FDA’s recent guidance, sponsors have been given more clarity on a rigorous pathway to use EHR derived data for both pre- and post-market submissions, and that has really set the bar in terms of the expectation and requirements for demonstrable data provenance, quality checks, transparent transformation, processes and pipelines, as well as understanding what data is fit for purpose for specific regulatory uses.
EHR-derived evidence is therefore no longer just supportive; it’s becoming more strategic and more integrated upstream in the development and commercialisation process of these treatments.
This latest guidance is a step in the right direction, stating that the FDA is open to considering deidentified patient data as part of regulatory submissions. The ability to use that, as opposed to pseudonymised, traditional clinical study type data, really encourages sponsors to consider the vast number of deidentified data sources that are available for research.
What it doesn’t say, however, is equally important. It doesn’t say that there’s going to be any sacrifice on the bar of data quality and other standards that data needs to meet for it to be used for regulatory submission.
AB: What changes do you anticipate from sponsors based on this clarity? Do you expect a higher uptake of alternative trial designs to utilise this data?
JL: There are always considerations that, with all the secondary data and innovation available, do we replace the sort of legacy way of doing things? I think the answer continues to be no.
EHR-based RWE, as with any other secondary database, is never going to replace randomised, controlled trials, but I see it as a way to fill a gap that sponsors can’t practically cover and enabling more flexibility to have smarter and more efficient right trial design.
Using it for external or hybrid controls, additional context arms, or different context terms that might have been generated historically via a prospective registry or natural history study is possible. It could also be used to support label expansions and subpopulations. It means sponsors are better able to understand the landscape of patients receiving treatment and the outcomes that they are experiencing, and the ability to again track those passively.
Finally, the flexibility in not just operationalising, but also designing and feeling confident in the design of adaptive and hybrid studies, which is more possible than ever before, makes it exciting and an innovative environment in the trial space.
AB: Are there any problems with EHR data that will make it difficult for sponsors to integrate it into studies?
JL: All of us who have worked in RWD, especially in a regulatory context, have been humbled by RWD before. The main pain points are missing data, bias in the way that the data are collected, and the need to define and validate endpoints, particularly endpoints that sponsors would have designed into a trial. With RWD, you need to look at these retrospectively.
However, there are avoidable pain points, like data standards and traceability. A sponsor needs to show how EHR data were sourced and transformed into an analysis-ready data set and the documentation for it in a transparent way, as per the FDA guidelines. When we think about primary EHR data sources and collecting that information, sponsors and their supporting research organisations have a lot of control over doing that, but when it comes to secondary EHR data sources, they have to rely on the provider of that secondary data source for this.
EHR and RWD sources can be classified into two camps. You have sources that control and process themselves, the originators of the primary data, and then they do the transformation to deidentify it and make it research-ready. Then there are other secondary data sources that take that primary data from a source and aggregate it, which adds a level of obfuscation on where the data came from. Sometimes the aggregators will bring multiple primary sources together, so the more a sponsor moves away from the primary collection, the harder it becomes to articulate and the more complicated it gets.
AB: Is this something that has been replicated by other agencies, such as the EMA or MHRA?
JL: Other regulatory bodies are becoming more conscious of the need to understand the impact of new treatments in their own patient populations, whether that is pre-approval as evidence to support that approval application, or post-approval to understand the longer term.
There are different ways that RWD exists outside of the US, right from single payer healthcare systems that have longitudinal data on a patient, from cradle to grave, to systems that are more like the US, which are more fragmented. That secondary data context varies pretty dramatically country by country, so you end up with different ways to be able to provide RWE.
Overall, there is a very positive trend for RWD globally, not only because it gives you a more complete picture of that patient than what can be collected in trials, but it also has the potential to make studies faster, more efficient and more representative.
AB: From a sponsor’s perspective, what red flags could create issues during FDA submission?
JL: The best practice is ensuring transparency and communication and making sure that you have de-risked. Using RWD adds complexity to a process that we all know very well, and as this space evolves, sponsors are going to invariably run into bumps in the road along the way, but those things are always good to keep in the front of the mind.
With respect to the data, they must question whether they are working with the right partners to be able to take that data and make it suitable for the use case, whether that be part of the submission package or to inform the FDA for perspective on the package.
If you’re using RWD as part of the evidence package, then you need to treat it like you would traditional clinical trial data, but that can be harder to achieve with RWD because of how messy it can be. This is a challenge as you’re taking traditional principles and applying them to a new data source, so the methodology cannot be the same. With clinical trial data, you can go back to the source and validate but with RWD, you may not have this.
Despite all this, these motions and these learnings are incredibly useful, but they’re also strategically important for sponsors to get right, and this is what is going to ultimately prepare sponsors for the next wave of trial innovation and create smarter and more efficient studies.
